A Functional Joint Model for Survival and Multivariate Sparse Functional Data in Multi-Cohort Alzheimer's Disease...

Area of Science:

• Biostatistics
• Neurodegenerative Diseases
• Longitudinal Data Analysis

Background:

• Alzheimer's disease (AD) research often involves complex, multi-study data with missing outcomes.
• Existing models may not adequately integrate longitudinal health data with survival information, limiting comprehensive analysis.

Purpose of the Study:

• To develop an integrative joint model for analyzing multivariate sparse functional and survival data in Alzheimer's disease (AD) across multiple studies.
• To extend the multivariate functional mixed model (MFMM) to handle missing-by-design outcomes in multi-cohort studies.

Main Methods:

• Developed an extended multivariate functional mixed model (MFMM) integrating longitudinal outcomes and time-to-event data.
• Employed a parsimonious survival model to link disease progression trajectories to survival outcomes.
• Utilized penalized splines within an Expectation-Maximization (EM) algorithm for efficient parameter estimation.

Main Results:

• The model successfully captured shared disease progression trajectories and accounted for inter-cohort variability in Alzheimer's disease.
• Application to three AD cohorts demonstrated the model's ability to integrate diverse data types effectively.
• Simulation studies confirmed the robustness and accuracy of the proposed statistical framework.

Conclusions:

• The integrative joint model provides a flexible and interpretable framework for analyzing complex Alzheimer's disease data across multiple studies.
• This approach enhances the understanding of AD progression and supports clinical decision-making in multi-cohort research settings.
• The model's ability to handle sparse functional and survival data makes it valuable for future neurodegenerative disease research.